An Artificial Neural Network for Solar Power Generation Forecasting Using Weather Parameters

Abstract

According to the present context, electrical power generation of Sri Lanka primarily depends on hydro and thermal power plants. As a developing country with increasing electricity demand and strong national environmental policy, the focuses have been driven towards renewable power sources like wind and solar. As a result, number of wind and solar power projects in Sri Lanka has been encountering a considerable growth. Intermittency in the solar Photovoltaic (PV) power generation can significantly increase the variations in the supply side, especially when the solar power penetration is high. Accurate forecasting of solar power generation helps system control engineers with effective and efficient power plant dispatching and scheduling. Weather parameters such as solar irradiance, cloud cover and wind speed determine the solar power output of a PV panel. Machine learning methods such as neural networks, support vector machines and regression models have shown high performance on time series forecasting. In this paper, an Artificial Neural Network (ANN) is proposed to predict solar power generation using weather parameters. An application study is conducted using the Buruthakanda solar park. The results show that the forecasting performance of the proposed ANN model outruns the Smart Persistence (SP) model.